A bilingual research brief for The Mid-2026 Semiconductor Correction: AI Infrastructure, Model Mis-Specification, and Memory as a Bottleneck Asset — SSRN Working Paper No. 7013899, Huang, Po-Sung (Sinclair), 2026.
What the paper studies. The mid-2026 semiconductor correction generated two competing narratives: a “bubble-burst” reading grounded in legacy DRAM cycle models, and an “infrastructure repricing” reading grounded in bottleneck economics and AI capex fundamentals. The paper uses the divergence in analyst and retail commentary around Micron Technology as a case study in regime mis-specification — the application of a prior-cycle statistical model to a structurally transformed market environment.
The framework. Its primary independent contribution is a three-dimensional analytical instrument: value-chain positioning via the Irreplaceability Index (IRI); cash-flow time structure via Equipment-CapEx Divergence Ratio (ECDR) context; and gross margin as a real-time regime indicator. Together they are offered as a replicable way to distinguish durable AI-infrastructure premiums from speculative excess under market stress.
What it concludes. The June 2026 correction is broadly consistent with a macrofinancial repricing event rather than a collapse in AI infrastructure demand. The paper is a stress test of the framework under live market conditions — not a price forecast.
Where it sits in this research. This is the featured core paper of the AI Infrastructure reading path, where the same IRI/ECDR/gross-margin lens connects the capex money map, deployment constraints, and bandwidth economics essays.
中文｜核心觀點
2026 年中的半導體修正出現兩種對立敘事：以傳統 DRAM 週期模型為基礎的「泡沫破裂」解讀，與以瓶頸經濟學、AI 資本支出基本面為基礎的「基礎設施重定價」解讀。本文以美光（Micron）為個案，檢視分析師與散戶評論的分歧，並將其定性為「制度誤設（regime mis-specification）」——把上一個週期的統計模型，套用在結構已經轉變的市場上。
論文的核心貢獻是一組三維分析框架：以不可替代性指數（IRI）判斷價值鏈位置、以設備－資本支出背離比（ECDR）脈絡判讀現金流的時間結構、以毛利率作為即時的制度指標。結論認為，2026 年 6 月的修正大體上符合總體金融面的重定價，而非 AI 基礎設施需求的崩壞；本文是對框架的壓力測試，不是價格預測。